Retrospective clinical reports have been the source of most of the information about AZT myopathy. Due to a lack of basic scientific studies, the pathophysiology of AZT myopathy is unknown. In order to better understand this disorder, we propose to combine epidemiologic and clinical methods with in vivo and in vitro modeling to test hypotheses about the mechanism of AZT toxicity. Our hypotheses include: defective mitochondrial biogenesis; abnormal mitochondrial energy production; membrane damage, inducing recurrent myonecrosis; and lysosomal damage. First, we propose prospective epidemiologic studies to determine relative incidence and risk factors for AZT myopathy and clinical studies to obtain evidence of myotoxicity from longitudinal drug challenge/dechallenge/rechallenge data. Two unique sources will be studied-subjects from well-defined longitudinal cohorts of AZT-using and control HIV-positive gay men (MACS/SHARE) and IV drug users (ALIVE, which includes minorities)- as well as patients from our large community referral networks. The use of cohorts allows precise statements about incidence of toxic myopathy and identifies its risk factors (dose, renal or hepatic insufficiency, concomitant drug use). Using pathologic material, we plan to develop reliable pathologic criteria for drug induced myopathy, as distinguished from HIV myopathy. Second, we propose animal modeling. In preliminary studies, we have been able to model myopathic changes in mice that are intoxicated chronically with AZT. This material will allow an analysis of the earliest morphologic events in vivo (e.g., mitochondrial vs lysosomal vs necrosis). It will allow comparisons with other toxicities (e.g., anemia) and experimental manipulation of other risk factors (e.g., other drugs metabolized by glucuronidation). A third approach is in vitro analysis of the mechanisms of AZT toxicity in cultured skeletal muscle cells. This allows precise control of conditions surrounding toxicity, observation of the earliest ultrastructural changes, and analysis of molecular events-muscle cell growth (morphology and cell creatine kinase levels), mitochondrial function (lactate production), mitochondrial proliferation and turnover (mitochondrial DNA quantification), and muscle sarcolemmal leakage and damage (creatine kinase release). The time course of all of these pathologic changes will be correlated with measured changes in AZT-5'-triphosphate, the active metabolite of the drug. Lastly, the mechanism of AZT toxicity in these paradigms will be compared and contrasted with 2',3'-dideoxycytidine, which could cause myotoxicity but is not yet known to cause human myopathy. These experiments offer the chance to study toxic mechanisms in a clinically-relevant tissue-skeletal muscle-which may have broader relevance to other tissues as well. AZT myopathy is common in and of itself and is of immense importance to the treating physician faced with the decision to stop an essential, proven, but unique therapy. A full understanding of the mechanisms of myotoxicity may allow us to prevent or control this complication, to design analogues free of toxicity, or to better predict and test for toxicity in other newly developed nucleoside analogues.